The organizers invited 23 international carbohydrate scientists (including mass spectrometrists, NMR spectroscopists, and computer database experts) to attend a workshop during August 18-20, 1995, directly preceding the XIII International Symposium on Glycoconjugates in Seattle, Washington. The workshop, sponsored by the U. S. Department of Energy, was organized by the CCRC and the Department of Bio-Organic Chemistry of the University of Utrecht (The Netherlands). The workshop met to formulate guidelines for establishing spectral databases of complex carbohydrates. Elucidating the functions of complex carbohydrates depends on knowledge of their structures. There has been a great increase in the level of interest in the structures and functions of complex carbohydrates. Much of this interest stems from the increasing recognition that the asparagine-linked (N-linked) and serine-or threonine-linked (O-linked) carbohydrate chains of glycoproteins, the carbohydrate chains of glycolipids, and oligosaccharide fragments of glycosaminoglycans are involved in many of the most important diseases of man, including autoimmune diseases, inflammatory diseases, cancer, and inherited storage diseases. Complex carbohydrates are also strategically involved in many of the normal physiological functions and developmental processes of the cells of all organisms. The databases will enable the scientific community to avoid the waste of research effort and funds that frequently occurs when carbohydrate chemists are forced to duplicate the structural characterization of previously characterized complex carbohydrates. In the absence of spectral databases they are unaware they are analyzing a known molecule until they have completely determined its structure. When the collections of mass and of nuclear magnetic resonance (NMR) spectra initiated at the workshop are subsequently developed into searchable databases, scientists only need query the databases with the spectrum or with information defining the spectrum of an unidentified carbohydrate to find out if it has been previously characterized. The non-carbohydrate specialist will also benefit from the availability of easily searchable spectral databases. The following are among the unanimous recommendations reached by the workshop participants. (i) An Internet listserv will be organized to facilitate e-mail communications between those wishing to establish and utilize collections of carbohydrate spectra. William York has already established the CHO-DATA listserv. (ii) Carbohydrate spectral data collections will be organized by volunteer curators at FTP (File Transfer Protocol) sites on the Internet. The sites will be connected electronically to form an interactive distributed collection of carbohydrate spectra. (iii) All carbohydrate spectra will be made freely available to anyone who wishes to download them from the FTP site. (iv) Carbohydrate scientists will be urged to submit their published spectra to the appropriate FTP sites using standardized protocols available at each FTP site. The protocols will also be sent to all subscribers to the CHO-DATA listserv and anyone else so requesting. (v) The spectral collections will be associated with the Complex Carbohydrate Structure Database (CCSD) by attaching to each spectrum the appropriate CCSD substance identifier number. (vi) The collections of spectra will form the foundation for the formation of databases with easy-to-use search engines that utilize appropriate semantics and syntax. It is hoped that one or more databases will be formed from each collection of spectra. The databases may be formed by academic or commercial groups.